Re: Marantz RC5 22 bits (Extend Data Word), possible with lircd.conf semantic ?

[Mona <de...@ch...>]

lirc_ml_sf wrote
> The main of these two others is a RC5 modified to be longer up to 22 bits
> in 
> length (17 bits of data). It seems this is called Marantz Extend Data 
> Word (see chronogram below).
> 
> My question is : How to introduce two space bits (below in cyan) 
> inside an RC5 (that never returns to zero by definition) ?
> 
> bffebjjh.png (28K)
> http://lirc.10951.n7.nabble.com/attachment/9784/0/bffebjjh.png

Your Marantz Extend Data Word and chronogram is actually the RC5x protocol.
I have the same problem with a Marantz SR8000 AV receiver (RC2000 MkII or
RC-18SR remote). I almost got it to work using "pre" and "pre_data". For
example, Marantz code 16 12 01 (discrete power on) would encode as 1101 0000
(2 bit space) 0011 0000 0001.

begin remote
  name        RC5x
  bits           20
  flags         RC5|CONST_LENGTH
  eps           30
  aeps         100
  one          890   890
  zero          890   890
  gap          125000
  toggle_mask  0x20000   # toggle_bit 3
  frequency      36000

  pre_data_bits  8
  pre_data        0xD0       # System = 16, limited to commands 0-63
  pre                0    3560 # No pre pulse, pre space of 2 bit times
(890*4)

  begin codes
    power_on     0x301 # 16 12 01
    power_off     0x302 # 16 12 02
  end codes
end remote

Unfortunately, lircd won't send the pre pulse/space unless BOTH values are
non-zero. It would be an easy code fix that shouldn't break anything else.
>From lircd 0.9.0 (transmit.c):

  290  if (remote->pre_p > 0 && remote->pre_s > 0) {
  291     send_pulse(remote->pre_p);
  292     send_space(remote->pre_s);
  293  }

Change this to:

 if (remote->pre_p > 0) {
    send_pulse(remote->pre_p);
 }

 if (remote->pre_s > 0)  {
    send_space(remote->pre_s);
 }

LIRC coders: I would suggest similar changes to the send_post, send_header,
and send_foot functions. It would increase flexibility without breaking
existing lircd.conf files. Or skip all the zero checks and have the
send_pulse and send_space functions immediately exit if called with data ==
0 (although this might have other side effects, mostly with invalid
lircd.conf files and the check_send_buffer function).

For now I'm using the raw_codes option. The RC5x commands I've needed are
discrete commands, so not having the toggle bit hasn't been a problem. Raw
codes for the commands above can be seen  here
<http://lirc.10951.n7.nabble.com/trouble-with-rc5x-td5701.html>  .

Adrian




--
View this message in context: http://lirc.10951.n7.nabble.com/Marantz-RC5-22-bits-Extend-Data-Word-possible-with-lircd-conf-semantic-tp9784p9788.html
Sent from the LIRC mailing list archive at Nabble.com.